Reply to comment by M. S. Bos on “Anomalous ocean load tide signal observed in lake‐level variations in Tierra del Fuego”

Abstract

[1] The comments of Bos [2010] are valuable and certainly contribute to an improved understanding of the tidal interactions in Tierra del Fuego. They are based on a thorough examination of the subject and emphasize the scientific relevance of this investigation. [2] Bos [2010, Figure 2] presents modelled lake-tide predictions that differ from those given in our article. The reason for this difference consists essentially in incorrect phase lags for the direct effect as given by Richter et al. [2009, Figure 2a]. Due to inconsistencies between the used ETERNA and TASK-2000 software tools an unnoticed time reference shift of one hour was introduced which resulted in the about 30° too large phase lags. [3] However, our modelled load tides (indirect effect) are not affected thereby. Still, the load tides predicted by Bos [2010] are about 10% larger in amplitude compared to our results. This is most likely explained by the application of different load models for the Magellan Strait (e.g., our Magellan Strait model produces a slight reduction in the total load tide amplitude in contrast to the statement by Bos). An impact of accounting for the complete global tidal load model rather than our limited domain in the order of 0.2 mm, 15° as stated by Bos seems to us questionable. Test computations had shown that using the tidal loading predictions of M. S. Bos and H.-G. Scherneck (Ocean tide loading provider, 2009, available at http://www.oso.chalmers.se/∼loading/) (implying convolution over the complete global model; FES2004 ocean tide model, Gutenberg-Bullen earth model) changes the load tide contributions at the three tide gauge sites by no more than 0.01 mm and 1.5°, respectively (crustal deformation effect only). [4] A corrected version of Richter et al.'s [2009] Figure 2a is displayed in Figure 1. It reveals now a smaller difference between observed and modelled lake tide signals. However, the remaining misclosure still exceeds the uncertainties of the observed signal and also the 3σ interval (which has been added in Figure 1). Here, σ represents the site-specific standard deviation of the M2 amplitude and M2 phase as determined from a 12-months lake-level observation record. For site B, the σ were derived directly as the standard deviation of the M2 tidal parameters of three independent 12-month records (σ12B). For sites A and C, for which only one year of observations is available, the respective σ were derived from the standard deviations of the M2 parameters of two independent 6-month records, scaled by the ratio σ12B/σ6B between the standard deviations of the 12-month solution and two 6-month solutions (σ6B) at site B (assuming that the gain in accuracy by increasing the record length from 6 to 12 months is equal at all three sites). The geometrical relation between the vectors (i.e., relative phase relation) continues to suggest that this significant difference between observation and model is to be attributed to an underestimation of the modelled load tide amplitudes. A satisfactory agreement would require a scaling of the modelled load tide amplitudes by about 120%. [5] The article was aimed to present our lake-tide observations and their deviation from the model predictions. Our considerations concerning the elastic earth model were only one general interpretation/speculation in an attempt to explain the obtained differences. We did, however, not pretend to infer a realistic elastic earth model from the tidal parameters. For such an undertaking our approach (constant scaling of the complete Green's function, no consideration of vertical crustal structure, etc.) would indeed be too simplistic. Therefore, the discussion of this topic by Bos [2010, Figure 1] represents a valuable extension of knowledge. [6] Thus, thanks to the comment by Bos [2010] the deviation of the modelled load tide from the observations has been reduced from 30% to 20%. Nevertheless, a 20% amplification of the theoretical load tides is still important compared to other published observational results. And, in our opinion, it is still difficult to explain it by uncertainties of the observations and the load (ocean tide) model only. On the other hand, a substantial amplification of elastic loading effects has been observed also in other regions in South America [Bevis et al., 2004]. However, we agree with Bos that the differences between the observations in Lago Fagnano and the predictions of the internet ocean tide loading provider should be interpreted with care. At present, the authors are continuing the investigations of tidal interactions in Tierra del Fuego and the results will be made public to the scientific community in due time.